FSO systems present many advantages like high data rate, license-free bandwidth and tap-proof communication allowing the download of vast amounts of data from LEO satellites. However, the atmospheric channel is quite challenging, because of spurious effects such as absorption, scattering and scintillation that in turn vary the link losses in correspondence to the elevation. In order to maximize the downlink throughput, it should start around 5° elevation. This leads to a design with suboptimal performance for higher elevations when constant data rates are used. Therefore, DLR is developing a system to adjust the data rate according to elevation and atmospheric channel conditions. This data rate variation is achieved by determining the maximum rate for higher elevation and then for lowering the data rates a bit-level repetition is performed. The presented system enables a fast transition between the different data rates. Additionally, this system allows the satellite to transmit data at rates even lower than those nominally supported by the physical transceiver. At the receiver side, the system complexity increases as it should be able to acquire, detect, and filter the signal for different data rates. DLR proposes a system that mirrors the operation of its transmitter counterpart by sampling the acquired signal at the maximum data rate. Then an FPGA processes the signal by majority decision algorithm followed by voting system that filters and detects the intended data rate in real-time. This enables replication and parallelization of the filtering and detection processes enabling the automatic detection of the received data rate. In order to provide noise stability, the transition between data rates is governed by a hysteresis process. This scheme allows the detection and selection of the proper data rate in the range of few microseconds for a system operating between 10 Gbps and 1.25 Gbps in steps of factor of 2, ignoring the propagation delays.